NPC2 is expressed in human and murine liver and secreted into bile: Potential implications for body cholesterol homeostasis


  • Potential conflict of interest: Nothing to report.


The liver plays a critical role in the metabolism of lipoprotein cholesterol and in controlling its elimination through the bile. Niemann-Pick type C 2 (NPC2), a cholesterol-binding protein, is key for normal intracellular trafficking of lipoprotein cholesterol, allowing its exit from the endolysosomal pathway into the metabolically active pool of the cell. In addition, NPC2 is a secretory protein from astrocytes and epididymal cells. Although NPC2 mRNA is detected in the liver, plasma and biliary NPC2 protein levels and function have not been reported. This study demonstrates that NPC2 is present in murine and human plasma and bile. In addition, hepatic NPC2 protein expression was dramatically increased in NPC1-deficient mice but not regulated by cholesterol feeding or pharmacological modulation of various nuclear receptors involved in cholesterol and bile acid metabolism. Interestingly, biliary NPC2 levels were 3-fold increased in gallstone-susceptible C57BL6/J versus gallstone-resistant BALB/c mice. Furthermore, NPC2 was exclusively found in the cholesterol pro-nucleating ConA-binding fraction of human bile. In conclusion, NPC2 is secreted from the liver into bile and plasma, where it may have a functional role in cholesterol transport in normal and disease conditions. (HEPATOLOGY 2006;43:126–133.)

Niemann Pick type C (NPC) disease is a fatal, autosomal recessive lipid storage disorder characterized by cholesterol accumulation in the liver, spleen, and central nervous system.1, 2 Mutations in NPC1 and NPC2 genes are responsible for this disease condition.3, 4 NPC1 encodes an intracellular transmembrane protein with a putative sterol sensing domain similar to those found in hydroxymethylglutaryl-coenzyme A reductase and sterol regulatory element binding protein cleavage-activating protein,3, 5 which are known to play important roles in maintaining cellular sterol balance. Conversely, NPC2 encodes a small soluble lysosomal cholesterol-binding protein.4, 6–8 Thus, the pathogenesis of NPC disease is closely related to abnormal cellular cholesterol transport and deposition.

NPC proteins appear to play important roles in intracellular traffic of both endogenously synthesized cholesterol and low-density lipoprotein (LDL)–derived cholesterol acquired by endocytosis.1, 2, 9 NPC1 overexpression in CHO cells increased cholesterol synthesis and total cellular cholesterol concentration.10 Cultured cells harboring mutations in NPC1 or NPC2 genes accumulate endocytosed lipoprotein cholesterol within lysosomes and exhibit delayed sterol-regulated gene expression.1, 2, 11, 12 More recently, NPC1 and NPC2 were found to regulate cellular cholesterol homeostasis through generation of LDL cholesterol–derived oxysterols.12 Taken together, these studies demonstrate that NPC1 and NPC2 play critical roles in regulating intracellular cholesterol transport and homeostasis.

Interestingly, NPC proteins also may modulate overall body cholesterol metabolism. We recently found that NPC1 plays a key role in controlling the transhepatic trafficking of cholesterol from plasma into the bile in mice.13 Also, the studies of Xie et al.14–16 have shown that cholesterol carried in LDL and chylomicrons was sequestered in an intracellular pool of a variety of tissues in NPC1-deficient mice [NPC1(−/−) mice].14–16 In contrast, the significance of NPC2 for in vivo cholesterol homeostasis has not been fully established. NPC2 is a major protein of epididymal fluid, where it may modulate sperm formation.6, 17 However, NPC2 expression has been detected in several tissues, including neurons and astrocytes,18, 19 suggesting that it could have a more global function in lipid homeostasis.4, 17

In mammals, the liver plays a critical role in lipoprotein cholesterol metabolism and is a key organ for body cholesterol removal into the bile.20, 21 Hepatocytes acquire cholesterol by 3 metabolic pathways: (1) endogenous biosynthesis from acetate, (2) receptor-mediated endocytosis of chylomicrons, very low-density lipoproteins and LDL, and (3) selective cholesterol uptake from high-density lipoproteins via the scavenger receptor class B, type I.21 The receptor-mediated endocytic pathway is one of the major mechanisms for uptake of lipoprotein cholesterol into the liver. In fact, more than 80% of circulating plasma LDL cholesterol is cleared by endocytosis in this organ.20 Furthermore, the hepatic endocytic pathway is also responsible for metabolism of lipoprotein remnants. Finally, hepatocytes efficiently eliminate sterols through the bile as unesterified cholesterol and by its catabolism and biliary secretion as bile acids.21 Biliary cholesterol disposal is critical not only for normal body cholesterol homeostasis, but also for the pathogenesis of cholesterol gallstone disease, a highly prevalent and costly condition in Western countries.22–24

We have established that hepatic NPC1 is critical in controlling plasma cholesterol and biliary lipid secretion in the murine liver.13 However, the expression and functional relevance of NPC2 in the hepatocytes are not fully understood. We evaluated the expression and regulation of NPC2 in the murine liver as well as the consequence of adenovirus-mediated hepatic NPC2 overexpression in mice. We also analyzed the presence of NPC2 protein in plasma and bile of human and 2 murine strains with different susceptibility to diet-induced gallstone formation.


NPC, Niemann-Pick type C; NPC1, Niemann-Pick C1 protein; NPC2, Niemann-Pick C2 protein; LDL, low-density lipoprotein; NPC1 (−/−) mice, NPC1-deficient BALB/c mice; wild-type mice,control BALB/c mice; Con-A, concanavalin–A; PNGaseF, peptide N-glycosidase F.

Materials and Methods

Animals and Diets.

The C57BL/6J murine strain was originally purchased from Jackson Laboratory (Bar Harbor, ME) and bred to generate our own colony. BALB/c mice carrying a heterozygous mutation in the NPC1 gene5 were donated by Dr. Peter Pentchev from the National Institutes of Health (NIH; Bethesda, MD) and were used to generate animals that were wild-type [NPC1 (+/+)] controls and homozygous NPC1(−/−) mutants. The genotype of offsprings from NPC1 (+/−) crosses was identified using a polymerase chain reaction–based screening as previously described.5 All mice were maintained with free access to water and a chow diet (<0.02% cholesterol; Prolab RMH 3000, PMI Feeds Inc., St. Louis, MO).

Wild-type male C57BL/6J and BALB/c mice (2-3 months old) were used in experimental protocols. In some experiments, C57BL/6J mice (2-3 months old) were switched from chow diet to diets containing 2% cholesterol (Sigma, St. Louis, MO) or 0.2% ciprofibrate (Sanofi, Gentilly, France) for 7 days, or 0.5% chenodeoxycholate (CDCA) (Sigma) for 48 hours. In other experiments, C57BL6/J mice were treated through oral gavage with either vehicle (0.5% methyl cellulose, Sigma) or 2.5 mg/mouse guggulsterone (Steraloids Inc., Newport, RI) or 1 mg/25 g mouse T0901317 (Amgen Inc., South San Francisco, CA) daily for 7 days. In one experiment, male BALB/c NPC1(−/−) mice of 6 weeks of age as well as age- and sex-matched BALB/c controls were used to compare hepatic NPC2 expression levels.

Protocols were performed according to accepted criteria for humane care of experimental animals and approved by the review board for animal studies of our institution.

Recombinant Adenoviruses Preparation and Administration.

A recombinant adenovirus containing the murine NPC2 cDNA (Ad.NPC2) under control of the cytomegalovirus promoter was generated by homologous recombination in bacterial cells using the AdEasy system generously provided by Dr. Bert Vogelstein (Johns Hopkins University, Baltimore, MD).25 The control adenovirus Ad.E1Δ without transgene was kindly donated by Dr. Karen Kozarsky (SmithKline Beecham Pharmaceuticals, King of Prussia, PA). Large-scale production of recombinant adenoviruses was performed in HEK293 cells as previously described.13

For in vivo administration of viruses, mice were anesthetized by ether inhalation, the femoral vein was exposed, and 1 × 1011 viral particles (in 0.1 mL isotonic saline buffer) of control or recombinant adenoviruses were injected intravenously. An additional control group received 0.1 mL saline buffer only. Mice were euthanized for analysis 7 days after adenoviral infection.

Murine Bile, Blood, Epididymal Fluid, and Liver Sampling.

Mice were anesthetized by intraperitoneal injection of sodium pentobarbital. The cyst duct was ligated and a common bile duct fistula was performed using a polyethylene catheter. Hepatic bile specimens followed by plasma and liver samples were obtained as described previously.26, 27 Mouse epididymis fluid was prepared as previously described.28

Human Plasma and Bile Sampling and Processing.

Human plasma was obtained from healthy volunteers. Hepatic bile samples were obtained from patients with indwelling T-tubes 3 to 5 days after cholecystectomy for gallstones and immediately delipidated as previously described.29, 30 Gallbladder bile was obtained during laparoscopic cholecystectomy. Biliary glycoproteins were extracted with Concanavalin–A (Con-A) sepharose beads using a method previously described.31

Plasma Biochemical Analyses.

Serum alkaline phosphatase and alanine aminotransferase were measured by routine automated methods in the clinical facilities of our institution.

Glycosidase Digestion and Immunoblotting Analysis.

For glycosidase digestion, liver homogenates were prepared as previously described.26 Sixty micrograms total protein were incubated with or without peptide N-glycosidase F (PNGaseF) (New England Biolabs, Ipswich, MA) according to manufacturer's instructions.

For hepatic NPC2 protein analysis, proteins (40-60 μg protein/sample from liver extracts with or without PNGaseF treatment; 10 μg protein/sample from epididymal fluid; 5 or 10 μL murine hepatic bile; 40 μg protein/sample of delipidated human hepatic or gallbladder bile; 2-3 μL plasma) were separated by 12% to 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotted using an anti-NPC2 antiserum donated by Dr. Peter Lobel (Robert Wood Johnson Medical School, Piscataway, NJ). An anti-albumin antibody was used as protein loading control for immunoblotting. For biliary cathepsin D and β-galactosidase protein analysis, 10 μL murine hepatic bile were separated by 10% sodium dodecyl sulfate polyacrilamide gel electorphoresis and immunoblotted using anti-Cathepsin D (Santa Cruz Biotechnology, Santa Cruz, CA) and anti–β-galactosidase (Promega, Madison, WI) antibodies. Antibody binding to protein samples was visualized using enhanced chemiluminescence and measured using the GS-525 Molecular Image System (Bio-Rad, Hercules, CA).

Hepatic Immunofluorescence.

Fresh-frozen livers were sectioned, and cryosections (4-5 μm) were fixed in acetone, rinsed 3 times in phosphate-buffered saline, permeabilized with 0.1% Triton X-100 for 15 minutes, blocked overnight in 10% goat serum in phosphate-buffered saline, and incubated for 2 hours at 37°C with the polyclonal antibody against NPC2 (1:100). As secondary antibody, fluorescein isothiocyanate–conjugated goat anti-rabbit IgG (dilution 1:150) was used. After washing in phosphate-buffered saline, samples were mounted with coverslips using Fluoromont-G (EMS, Fort Washington, MD). Labeled sections were examined by immunofluorescence microscopy.

Quantitative Northern Blot Analysis.

cDNA probes for murine NPC2 and 18S rRNA were prepared from total liver RNA by a standard reverse transcription polymerase chain reaction procedure using primers based on mouse cDNA sequences available through GeneBank databases.

For Northern blotting, total RNA was prepared from murine livers using the acid guanidinium thiocyanate-phenol-chloroform method.32 Equal amounts of total RNA (30 μg per individual mouse from each experimental group) were size-fractionated by agarose-formaldehyde gel electrophoresis and transferred to nylon membranes. Probes were labeled by the random primer method (Promega, Madison, WI) and used for hybridization as previously described.13 Radioactive bands were quantified by phosphorimaging using the GS-525 Molecular Image System (Bio-Rad). Results were normalized to the signal generated from hybridization of a [32P]-labeled mouse 18S rRNA probe on the same filter.


The statistical significance of the differences between the means of the experimental groups was evaluated using the Student t test for unpaired data. A difference was considered statistically significant at a P value < .05.


NPC2 Protein Expression and Regulation in the Murine Liver.

Using immunoblot analysis, NPC2 was detected in total liver extracts as protein bands of approximately 16 to 22 kd in BALB/c and C57BL6/J mice (Fig. 1A). This signal matches the molecular size of NPC2 protein present in murine epididymal fluid, which was used as a positive control for immunoblotting (Fig. 1A). We did not observe differences in the constitutive expression levels of NPC2 in the liver samples of these two mouse strains. NPC2 protein heterogeneity seems to be attributable to differences in posttranslational glycosylation. After glycosidase F treatment, NPC2 was visualized as a single immunoreactive band of faster mobility approximately 16 kd in the epididymal fluid and liver extracts from both BALB/c and C57BL6/J mice (Fig. 1B).

Figure 1.

Hepatic NPC2 protein in BALB/c and C57BL6/J mice. (A) For NPC2 expression analysis, proteins from liver homogenates were size fractionated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, immunoblotted with anti-NPC2 antibodies, and subjected to densitometric analysis. Sixty micrograms liver homogenate proteins were analyzed, and 10 μg epididymal fluid proteins were used as NPC2 positive control (C+). Relative NPC2 expression levels are shown after normalization for pre-albumin as a protein loading control. (B) For endoglycanase digestion, homogenates from epididymal fluid (10 μg protein/sample) and livers (60 μg protein/sample) from BALB/c and C57BL6/J mice were incubated with peptide N-glycosidase F (PNGaseF), followed by Western blotting using anti-NPC2 antibodies.

To explore the regulation of hepatic NPC2 protein in vivo, we next tested the effect of a high-cholesterol diet as well as different agonists for nuclear receptors involved in the regulation of cholesterol and bile salt metabolism in the liver. C57BL6/J mice were fed with diets containing 2% cholesterol or the PPARα agonist cipofibrate, for 7 days, or the primary bile acid and FXR agonist chenodeoxycholic acid for 2 days. Also, mice were treated by gavage with the FXR antagonist guggulsterone or the LXR agonist T0901317 for 7 days. These treatments did not change hepatic NPC2 protein expression significantly, suggesting that this protein is not under the control of transcriptional factors that are critical for the regulation of cholesterol and bile acid synthesis and transport in the liver (Fig. 2).

Figure 2.

Hepatic NPC2 protein levels after cholesterol feeding and pharmacological modulation of nuclear receptors in C57BL6/J mice. Hepatic NPC2 protein levels were analyzed in C57BL6/J mice fed a 2% cholesterol-, 0.2% fibrate-, or 0.5% chenodeoxycholic acid–containing diets or treated by gavage with guggulsterone (2.5 mg/day) or T0901317 (1 mg/day). Liver samples were removed, and total homogenates were prepared. Proteins (40 μg per lane) were fractionated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred to nitrocellulose, and immunoblotted with anti-NPC2 and anti-prealbumin antibodies. NPC2 protein expression was normalized to the signal generated by the anti-prealbumin antibody on the same membrane. Results are representative of three independent experiments.

NPC2 has been described as being overexpressed in fibroblasts of NPC1 patients.33 To evaluate whether this regulation also occurs in hepatocytes in vivo, we analyzed NPC2 mRNA and protein levels in livers of NPC1-deficient mice. As shown in Fig. 3A, hepatic NPC2 protein levels were increased by 5-fold in the absence of a functional NPC1 protein in mice. This finding was paralleled with a 70% increase in NPC2 mRNA levels in NPC1(−/−) livers as shown by Northern blot analysis (Fig. 3B). These results suggest that the hepatic deficiency of NPC1 protein is sensed in liver cells by upregulating NPC2 gene expression and protein levels in vivo.

Figure 3.

Hepatic NPC2 expression analysis by immuno- and northern blotting in wild-type and NPC1(−/−) mice. (A) Total homogenate proteins (60 μg per lane) were fractionated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with anti-NPC2 and anti-prealbumin antibodies. NPC2 protein expression was normalized to the signal generated by the anti-prealbumin antibody on the same membrane. (B) Total hepatic RNA was prepared, electrophoresed, and transferred to nylon membranes. Gene expression was evaluated by RNA blot hybridization with32P-labeled cDNA probes. The mRNA expression data for npc2 gene is shown after normalization against 18S rRNA. The results shown in this figure are representative of 2 or 3 independent expression analyses. NPC1 (+/+): wild-type mice; NPC1 (−/−): NPC1-deficient mice. *indicates a significant difference (P < .05).

Adenovirus-Mediated Hepatic NPC2 Overexpression in Mice.

Infection of C57BL6/J mice with recombinant adenovirus encoding NPC2 markedly increased NPC2 protein levels in the liver (Fig. 4A). The expression of NPC2 in liver was also characterized by immunofluorescence microscopy (Fig. 4B). As shown in Fig. 4B, NPC2 staining displayed a punctuate cytoplasmic pattern consistent with NPC2 protein localization in intracellular vesicular compartments as previously shown.33 Unfortunately, adenovirus-mediated NPC2 overexpression induced a strong hepatic inflammatory response as evidenced by a dramatic increase on liver enzyme levels in plasma of Ad.NPC2-infected mice compared with animals infected with the control adenovirus (data not shown). For this reason, we did not further characterize the effect of hepatic NPC2 overexpression on plasma, liver, and biliary parameters related to cholesterol metabolism in vivo.

Figure 4.

Hepatic NPC2 expression in mice after infection with an NPC2 recombinant adenovirus. C57BL/6 mice were infected with murine npc2 recombinant adenovirus (Ad.NPC2) or a control adenovirus (Ad.E1Δ). Seven days after adenoviral infection, liver samples were collected for NPC2 immunoblotting (A) and immunofluorescence (B) analysis.

NPC2 Protein Levels in Murine and Human Plasma, Liver, and Bile.

Because NPC2 has been found as a major secretory product in the epididymal fluid, we analyzed whether NPC2 could be detected in plasma and bile. As shown in Fig. 5A, plasma from BALB/c and C57BL6/J mice contained polypeptides that were immunoreactive with the NPC2 antibody. Furthermore, NPC2 protein was also detected in bile obtained from these 2 murine strains (Fig. 5B). Noteworthy, biliary NPC2 levels were 3-fold higher in gallstone-susceptible C57BL6/J mice compared with the gallstone-resistant BALB/c strain. To analyze whether NPC2 secretion into bile was specific or correlated with the secretion of common lysosomal markers, we also performed immunoblot analysis of cathepsin D and β-galactosidase in bile. We found that both lysosomal markers were also increased in murine bile from C57BL/6 compared with BALB/c (Fig. 5B), suggesting that biliary secretion of NPC2 in mice is not a specific event and it occurs in conjunction with the other lysosomal hydrolases. In any case, these results indicate that NPC2 is a secretory protein present in murine plasma and bile. These findings were further supported by increased levels of plasma and biliary NPC2 protein levels found in mice with adenovirus-mediated hepatic NPC2 overexpression (Fig. 5C).

Figure 5.

NPC2 plasma and biliary levels in mice. For NPC2 expression analysis, proteins were size fractionated by SDS-PAGE and immunoblotted with anti-NPC2 antibodies. (A) Two microliters of murine plasma were analyzed. Two micrograms of epididymal fluid proteins were used as NPC2 positive control (C+). (B) Ten microliters murine hepatic bile were analyzed with antibodies against NPC2, pre-albumin, cathepsin-D, and β-galactosidase. (C) Two microliters plasma and 5 μL of bile from Ad.E1Δ and Ad.NPC2-infected mice were analyzed.

Finally, we tested whether NPC2 was also present in human plasma and bile. The rabbit polyclonal antisera raised against recombinant NPC2/HE14 detected immunoreactive bands of approximately 25 to 27 kd in human plasma as well as hepatic and gallbladder bile (Fig. 6, lanes 3, 7, 8). Next, extraction of biliary glycoproteins was performed by standard Con A-Sepharose affinity purification as previously described.31 As shown in Fig. 6, lane 9, human biliary NPC2 was completely recovered in the Con A–binding fraction as expected for glycoprotein features of NPC2.34 The antigens detected with the NPC2 antibody in human plasma and bile had a molecular size range similar to that observed for NPC2 immunoreactivity in human liver samples (Fig. 6, lane 5). Interestingly, human NPC2 (Fig. 6, lanes 3, 5, 7, and 8) exhibited a molecular size bigger than that found in murine epididymal fluid, plasma, liver, and bile (Fig. 6, lanes 1, 2, 4, 6). These murine versus human differences in the NPC2 protein size have been previously reported19 and are fully consistent with an extra consensus site for N-glycosylation present in human NPC2 compared with its murine ortholog.17, 34

Figure 6.

NPC2 plasma and biliary levels in humans. For NPC2 expression analysis, proteins were size fractionated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, and immunoblotted with anti-NPC2 antibodies. Lane 1, 10 μg mouse epididymal fluid; lane 2, 3 μL mouse plasma; lane 3, 3 μL of human plasma; lane 4; 40 μg mouse liver homogenate; lane 5, 40 μg human liver homogenate; lane 6, 10 μL mouse hepatic bile; lane 7, 40 μg human hepatic bile after delipidation; lane 8, 40 μg of human gallbladder bile after delipidation; lane 9, Con-A bound fraction (Con-ABF) from a human gallbladder bile sample; lane 10, Con-A unbound fraction (Con-AUBF) from a human gallbladder bile sample.


This study demonstrates that NPC2 is present in murine and human plasma and bile. In addition, NPC2 was isolated in the cholesterol pro-nucleating Con A–binding fraction of human bile. Furthermore, biliary NPC2 levels were 3-fold increased in gallstone-susceptible C57BL6/J mice compared with the gallstone-resistant BALB/c strain. Conversely, hepatic expression was dramatically increased in NPC1-deficient mice, but not regulated by cholesterol feeding or various nuclear receptor ligands involved in cholesterol and bile acid metabolism in vivo.

One significant finding of this study was the dramatic increase on NPC2 levels observed in livers of NPC1 mutant mice. Our results extend previous data obtained in fibroblast cells33 and astrocytes19 and murine cerebellum,35 indicating that NPC1 deficiency is also sensed in vivo in liver cells by upregulating NPC2 expression. This reciprocal co-regulation suggests that NPC1 and NPC2 function by complementation in a common pathway. This idea is also supported by recent findings studying NPC2 hypomorphic mice (0%-4% of residual protein expressed in different tissues) and NPC1/NPC2 double mutant animals.36 In this latter study, single NPC1 or NPC2 mutants as well as double mutants were similar in disease onset and progression, biochemical patterns of lipid accumulation, and pathology.

To explore the regulation of hepatic NPC2 protein in vivo, we tested the effect of a high-cholesterol diet as well as different agonists for the nuclear receptors FXR, LXR, and PPARα involved in the regulation of cholesterol and bile salt metabolism in the liver. Neither of these treatments changed significantly hepatic NPC2 expression at the protein level, suggesting that intrahepatic cholesterol content does not play a dominant role in controlling hepatic NPC2 expression in mice and that this protein is not under the control of transcriptional factors that are critical for the regulation of cholesterol and bile acid synthesis and transport in the liver.

NPC2, a protein with lysosomal localization when found within cells,33 binds cholesterol with high affi- nity.7, 8 The phenotype of NPC2 mutant cells and the knowledge of its structure and intracellular location has supported a working model in which NPC2 binds unesterified cholesterol released within the lysosomes after lipoprotein endocytosis and degradation, facilitating cholesterol export from this organelle through NPC1 or other proteins to various subcellular compartments.37 Remarkably, NPC2 also has been reported as an extracellular secretory product. In fact, NPC2 is particularly abundant in the epididymal fluid, where it has been suggested that this protein plays an important role in sperm formation by reducing cholesterol content in male spermatozoidal membranes,6 which seems critical for sperm capacitation.38 Furthermore, bovine NPC2 (also known as EPV20) is mainly present in the milk.39 More recently, NPC2 has been reported to also be secreted from primary astrocytes independently from the secretion of astrocyte-derived sterols.19

Our major finding is the demonstration that NPC2 is also found in plasma and bile. The principal sources and the functional role of plasma and biliary NPC2 and its relation with body cholesterol trafficking and homeostasis require further study. We were able to demonstrate that NPC2 hepatic overexpression in the murine liver correlated with a significant increment in plasma and biliary NPC2 protein levels, suggesting that hepatocytes actively secrete NPC2 through the basolateral and canalicular domains. The secretion of NPC2 is rather unique and not just a secondary effect of the overexpression induced by the recombinant adenoviral vector. In fact, we have not detected plasma or biliary secretion of other lipid transport proteins (e.g., sterol carrier protein-2, caveolin-1, or caveolin-2) when overexpressed in mice26, 40 (also unpublished personal observations). Based on our findings, the liver may represent a main source of plasma and biliary NPC2, even though we cannot exclude significant contribution of nonhepatic tissues to plasma NPC2 levels or bile duct epithelial cells to the biliary content of this protein. The detailed mechanism by which NPC2 is secreted into plasma and bile is not known. Our results indicated that NPC2 secretion into bile correlated with biliary levels of soluble lysosome proteins, suggesting that it may share the mechanism underlying lysosomal hydrolases secretion. Biliary secretion of these enzymes into bile occurs through a microtubule-dependent mechanism, suggesting that lysosomes follow an exocytotic pathway in which the luminal content is discharged into plasma or bile after fusion of the lysosomal membrane with the sinusoidal or canalicular membrane.41–43

Physiologically, NPC2 synthesis and secretion from the liver may play a role in cholesterol secretion from the liver into plasma and bile and, if so, in the maintenance of normal cholesterol metabolism. However, recent studies have indicated that NPC2 is not co-secreted with cellular sterols, and its overexpression cannot drive an increase in cellular cholesterol efflux from cultured astrocytes.19 Whether these latter findings are also valid to liver cells remains to be studied. Unfortunately, adenovirus-mediated hepatic NPC2 overexpression induced a strong hepatic inflammatory response in C57BL6/J mice. Infection of this murine strain with adenoviruses carrying a transgene usually causes some inflammation, which is comparable to that produced with the control adenovirus (Ad.E1Δ). Currently, we do not have an apparent explanation for the major hepatic inflammation observed in NPC2-overexpressing mice. Because severe liver inflammation leads to significant modifications in cholesterol metabolism,44 evaluation of the specific role of hepatic NPC2 expression in cholesterol traffic and homeostasis in this animal model of NPC2 overexpression was not possible.

That the primary pathogenic event for gallstone formation is biliary cholesterol hypersecretion followed by cholesterol crystallization and crystal growth in bile21, 45 has been established. A variety of biliary proteins can influence these early and critical steps during gallstone formation.46–48 Considering its cholesterol-binding activity, NPC2 secretion into bile led us to speculate that it may modulate biliary cholesterol precipitation. Interestingly, we found that biliary, but not hepatic, NPC2 content was significantly increased in C57BL/6 mice susceptible to diet-induced gallstone formation compared with BALB/c mice. Because NPC2 is secreted into bile, steady-state protein levels in the liver could not fully reflect NPC2 production rate. Further pulse-chase studies and mRNA expression analysis are required to elucidate whether hepatic NPC2 biosynthesis is different in these 2 mouse strains. Whether the strain-dependent difference in biliary NPC2 plays a role in the different susceptibility to gallstone disease also needs additional study.

Furthermore, we demonstrated that NPC2 present in human bile was fully recovered in a Con A–binding fraction, which has been characterized as a potent promoter of cholesterol crystallization in vitro.31, 49 The cholesterol pronucleating activity of this protein mixture has been well established, and a variety of proteins of this fraction have been partially identified.31, 49, 50 Exactly which of these biliary proteins is more relevant in cholesterol crystallization and how these glycoproteins facilitate cholesterol nucleation and crystallization remain unknown.50 Some of these glycoproteins may facilitate the proximity and fusion of cholesterol molecules carried in vesicular transporters enriched in cholesterol. NPC2 is a novel protein found in this biliary cholesterol pronucleating Con A–binding fraction and the first that specifically binds cholesterol. Interestingly, secreted NPC2 seems to be functional, retaining its cholesterol-binding activity.6 Therefore, biliary NPC2 may play a relevant role in the cholesterol crystallization defect present in gallstone patients.

In conclusion, we have described that NPC2 is present in plasma and bile, suggesting a potential role for this protein in cholesterol transport and gallstone disease in mammals. Further studies are required to elucidate the physiological and pathological relevance of our findings.


The authors thank Flavio Nervi for constant support and helpful discussions. They also thank Margrit Schwarz and Ricardo Moreno for help in obtaining T0901317 and mice epididymal fluid, respectively.